AI First with Adam and Andy

Andy Sack and Adam Brotman, with colleague Rose Kelly, tackle a question facing every leader: when AI makes it cheap and fast to build products, software, and creative, and just as easy for competitors to copy them, what stays defensible?

Brotman points to brand trust, the customer data that comes with a long relationship, and full-stack user experience that extends well beyond surface design. Sack adds community, distribution, and proprietary data, and argues that an underrated moat is human insight: the creativity and taste to decide where to aim the AI, especially as a swarm of small competitors can now copy fast. Brotman cautions that AI will keep improving, but brands will still own the customer relationship.

Asked what one thing a CEO should build to compound over three to five years, both hosts choose distribution, which Sack calls the precursor to data and faster learning loops, though insufficient alone. Sack’s broader advice: invest heavily in your own and your team’s AI use.

Andy Sack and Adam Brotman, joined by colleague Rose Kelly, follow up on a LinkedIn Live with Greg Gottesman and Alex Ray of Pioneer Square Labs, who say they are on track to evaluate 400 to 500 ideas a year, up from about 100 two years ago. The hosts explore what changes inside a company when trying, learning, and killing an idea becomes dramatically cheaper.

Sack argues that for startups more iterations and more failure is better, since the goal is finding a differentiated niche with cost-effective distribution, and that where you aim the agentic machine matters most. Brotman adds that agents can now produce 50 strategic plans overnight, which does not help, but 50 creative variants with autonomous ad testing does. Sack adds that attention is increasingly scarce as content, software, and one-person companies multiply.

For larger companies, the advice is to embrace iteration and failure, stop waiting for perfect data before building, and recognize that inaction carries the most negative ROI.

Andy Sack, Adam Brotman, and colleague Rose Kelly unpack a workflow that almost wasn't one. Preparing a board presentation on a complex new initiative, Brotman asked Claude Opus 5.5 to turn a PowerPoint into a video, then pushed it to match the production quality of two reference videos.

On its own initiative, the agent searched his email and meeting transcripts, ran external research, and worked for two hours, producing what he describes as a one-shot result. The hosts contrast this with a year ago, when the same task meant stitching together multiple tools scene by scene, and argue agents now take on the work of knowing how to use AI. Brotman adds that the more persistent and connected an agent is, with memory, data access, and knowledge of your goals, the more it anticipates what you need.

For executives, the shift is from managing how work gets done to defining what outcome they want, then steering and validating the result with their own taste and judgment, a principle the hosts explore in their book, Agents Inc.

Andy Sack and Adam Brotman, joined by Rose Kelly, open up an internal Forum3 leadership debate about whether their standing advice on enterprise AI tools is already out of date. The trigger is a new wave of personal-agent tools, Grokbot, Muse, and Instinct, that strip away chat histories, folders, and dashboards for one ongoing conversation with an AI teammate.

The hosts trace their long-standing advice: pick one frontier model, standardize the company on it, and go fast, building fluency and admin visibility. Brotman shares a hypothesis he floated internally: a simpler, role-and-goal interface might help more employees become proficient than a feature-heavy dashboard, even if it hides the complexity underneath. Kelly adds a distinction from Forum3's CTO, reframing the question: model, app, agent harness, and interface are separate layers, so a simple front end can still route through the same platforms. Andy offers himself as a data point, shifting from roughly 80 percent Claude to 80 percent Grokbot within a month.

For now, the hosts' advice to a mid-sized company hasn't changed: choose one frontier platform, standardize, and go. But they encourage CEOs to personally try the newer tools, as an early look at where enterprise interfaces may be heading.

Andy Sack and Adam Brotman bring in Rose Kelly to help business leaders make sense of a term suddenly everywhere in the AI press: slowdown. The conversation starts with what triggered it: a widely discussed post from Anthropic CEO Dario Amodei following the Hugging Face incident and a researcher departure that raised alarm about frontier model safety.

Brotman clarifies that Amodei's proposal, echoed publicly by Sam Altman, Elon Musk, and Demis Hassabis, calls for pacing future frontier model development and adding safety oversight and coordination, not slowing the compute or tools available to businesses today. Sack draws a line between two separate concerns: the existential debate over model safety and superintelligence, and the practical question of what a company should do with AI capability that already exists. Kelly presses the hosts on how leaders should talk to anxious teams, agreeing that CEOs need a clear point of view before they can reassure anyone else.

Their advice to executives is direct: this news should not change an AI strategy. Even if model development froze today, most organizations are years behind current tools, so leaders should keep pushing adoption, use AI responsibly, and treat safety concerns and business strategy as separate conversations.

 In this episode of AI First with Adam and Andy, Adam Brotman shares a client story: a finance executive responsible for a high-stakes, board-level cash-flow forecast could not keep pace with colleagues who had adopted frontier AI models, leaving him weeks behind on work others were turning around in hours.

Andy Sack and Adam break down what actually separates effective AI use from noise: selecting the right model and reasoning level, granting it the right access, and rigorously checking its work. They connect the moment to Forum3's core argument in Agents Inc., that most organizations can already run with at least half the headcount they carry today, and to the harder question underneath it: what leaders owe employees whose skills are becoming obsolete almost overnight.

For executives, the message is direct. AI fluency is no longer optional, and the choice between investing in someone's transition and replacing them is already being made across companies, often without any formal policy guiding it.

Nirit Cohen spent three decades at Intel and the past fifteen years studying the future of work, publishing on the topic years before it became a boardroom priority. In this conversation with Adam Brotman and Andy Sack, she examines what happens to entry-level jobs and organizational structure as companies begin handing tasks to AI agents.

Cohen challenges the idea that work can simply transition from people to AI, arguing that this underestimates the human judgment embedded in even routine-looking tasks. She distinguishes employability, executing work someone else designed, from entrepreneurship, knowing how to think with AI tools to create new value, and argues the latter is becoming the defining skill of the agentic era. The discussion also covers how AI is unbundling jobs into tasks that no longer map to traditional professions, and how the independent economy is becoming a training ground for these new capabilities.

For executives, the takeaway is that organizational redesign, not headcount reduction, is the real challenge: leaders need to rebuild how work is structured and how judgment is taught, not just decide which tasks to automate.

Andy Sack and Adam Brotman unpack Ghost GDP, a term from the Citrini report examined in their book, Agents Inc. The concept describes a scenario where AI agents perform work once done by people, letting companies keep producing goods and services, and even grow profits, without a matching rise in employment.

They explain why this breaks the usual link between strong GDP and low unemployment: output and profitability can climb even as fewer people are needed to generate them, leaving workers and consumers under pressure on both income and affordability. That tension feeds a broader debate the hosts return to often: whether AI creates enough new entrepreneurial opportunity, through solopreneurship and one- or two-person companies, to offset the jobs it displaces. Brotman notes a solo operator's company can take many forms, from a product business to an individual contributor freelancing with AI assistance.

For executives, the discussion signals that headline economic indicators may increasingly diverge from workforce and consumer health, and that reading productivity, GDP, and unemployment data will require new context in an AI-driven economy. Sack and Brotman frame solopreneurship, and the rise of small AI-assisted teams, as a force leaders will need to track and adapt to.

Forum3 co-founders Andy Sack and Adam Brotman explain why restaurant chains have become central to their agentic AI strategy. Drawing on themes from their book, Agents Inc., they argue that the agentic era is changing the nature of competition, and that restaurants offer a rare combination of stability and opportunity: the core product will not be disrupted by AI, but nearly everything surrounding it can be.

The discussion covers why chain restaurants generate both high transaction frequency and rich structured and unstructured data, from customer reviews and labor scheduling to pricing and marketing, creating ideal conditions for AI-driven analysis and decision-making. Brotman shares a concrete example: an AI agent producing a nightly research report on restaurant industry news, illustrating how quickly agentic workflows can be put into practice. Sack adds that gains from agentic improvements compound rapidly once applied across a chain's full footprint of locations.

For executives, the takeaway is that industries facing the least disruption to their core product may still see the most transformation in how they operate. Forum3's bet is that a narrow, defensible focus on restaurant chains, paired with agentic AI, positions both the company and its restaurant partners to compete as this shift accelerates.

What happens when executives no longer need to navigate complex software interfaces, or rely on specialists to retrieve the information they need? Winslow co-founder Niel Robertson joins Adam Brotman and Andy Sack to explore how AI agents and conversational interfaces are reshaping HR technology, SaaS platforms, and systems of record.

Robertson explains how agentic software can combine operational tools with domain knowledge, enabling leaders to manage payroll, employee data, onboarding, reporting, and compliance through natural-language requests. The conversation also examines the growing importance of operations roles as AI absorbs more clerical and execution-focused work across HR, sales, marketing, and other functions.

For executives, the implications extend beyond productivity. Organizations must reconsider data architecture, software investments, workforce design, employee privacy, and the role of human expertise. The result is a new operating model in which AI handles more routine work while people focus on judgment, strategy, change management, and organizational growth.